DeepSeek V2 Models — Hardware Requirements
5 DeepSeek V2 models from DeepSeek and the community, from the smallest that runs in 10.2 GB of VRAM at Q4 up to 235.7B parameters. Every row links to full quantization tables and GPU compatibility.
All DeepSeek V2 Models by Size
| Model | Params | Runs from (Q4) | Context | Publisher | Quant downloads |
|---|---|---|---|---|---|
| DeepSeek v2 Lite Chat | 15.7B | 10.2 GB | 164K | ||
| DeepSeek v2 Lite | 15.7B | 10.2 GB | 164K | ||
| DeepSeek V2.5 | 235.7B | 142.9 GB | 164K | ||
| DeepSeek V2.5 1210 | 235.7B | 142.9 GB | 164K | ||
| DeepSeek v2 | 235.7B | 142.9 GB | 164K | ||
| DeepSeek v2 Chat | 235.7B | 142.9 GB | 164K |
Frequently Asked Questions
- How much VRAM do I need to run a DeepSeek V2 model?
- The smallest DeepSeek V2 model, DeepSeek v2 Lite Chat, runs from 10.2 GB of VRAM at Q4_K_M (smaller low-bit quantizations start at 5.1 GB with a noticeable quality loss). Larger family members need proportionally more — see the table above for every model.
- Which DeepSeek V2 models can I run on a 16 GB GPU?
- 2 of 6 DeepSeek V2 models fit in 16 GB of VRAM at Q4_K_M, including DeepSeek v2 Lite Chat, DeepSeek v2 Lite.
- What is the most popular DeepSeek V2 model to run locally?
- DeepSeek v2 Lite Chat is the most downloaded DeepSeek V2 model in local-friendly quantized formats. It runs from 10.2 GB of VRAM at Q4_K_M.